PUENYAI Bot: A Retrieval-Augmented LLM Chatbot for University Information Services
Abstract
We present PUENYAI, a locally hosted retrieval-augmented generation (RAG) chatbot that answers enquiries on curricula, scholarships, and application procedures using institutional sources. Built entirely from open-source components and deployed on-premises, the system avoids external APIs while keeping answers grounded in official documents. This paper contributes three practical outcomes for higher-education deployments: (i) a reproducible on-prem blueprint—client, backend, embeddings, vector store, and locally served large language model (LLM)—together with the design trade-offs that led to these choices; (ii) a lightweight evaluation protocol (self-queries, structured user tests, and advisor review) that programs can reuse to check relevance, completeness, and policy alignment; and (iii) a set of lessons learned and governance guidelines (coverage of sources, chunking granularity, update workflow, and privacy controls) distilled from implementation. Results observed during formative testing indicate that a compact local stack consistently answers routine enquiries while reducing staff effort; remaining limitations (multilingual handling, admin tooling, and document structure) inform a practical roadmap for campus roll-outs.
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.